Operational Behavior Computing: Task Assignment Optimization Considering Worker Behavior and Algorithmic Responsibility

This study addresses a task assignment optimization problem by accounting for the heterogeneity of employee behavior and the accountability of algorithm deployment. Initially, we explore the variability within picker behavior across multiple dimensions, with experiments indicating notable differences. These behavioral variabilities are then integrated into a fairness-constrained task assignment model, aiming to minimize the maximum task completion time. The model's predictive accuracy is confirmed through some machine learning algorithms. A comparative analysis of various algorithms, including Variable Neighborhood Search (VNS), Adaptive Large Neighborhood Search (ALNS), Harmony Search (HS), Artificial Bee Colony (ABC), and Genetic Algorithm (GA), indicates that VNS and ALNS provide enhanced interpretability and timeliness, which are crucial for real-world scenarios demanding rapid response times, such as those in the second or millisecond range.